from datasets import load_dataset
from googletrans import Translator

def main():
    translator = Translator()
    raw_ds = load_dataset('SetFit/amazon_massive_intent_ja-JP')
    for split, dset in raw_ds.items():     
        label_text = dset['label_text']
        translations = []
        progress = 0
        for text in label_text:
            progress+=1
            if progress%1000 == 0:
                print(progress)
                print(translation)
            text = text.replace("_", " ")
            translation = translator.translate(text, dest='japanese').text
            translations.append(translation)
        dset = dset.add_column("label_text_jp", translations)        
        dset.to_json(f"{split}.jsonl")

if __name__ == "__main__":
	main()